Gemini Live Gets a Face: What the New Avatar Can Do

TVS News AI, editorially supervised, presents Gemini Live Gets a Face: What the New Avatar Can Do.

Imagine you're talking to your AI assistant – and it looks at you, nods, and reacts to your facial expressions. Google is now making this possible with Gemini Live.

Anyone who works with AI chatbots knows the often somewhat sterile feeling: you type or speak, and an invisible machine responds. With the new avatar, Google wants to make the dialogue more human – with a face that sees, speaks, and reacts to you.

According to reports by Golem and Heise, Google is expanding Gemini with animated avatars. The avatar speaks 97 languages according to Golem. Heise additionally describes that it can see and facially react to users.

But not everyone can use the feature yet: both sources report in agreement that the avatar is initially only available to business customers.

What can the new Gemini avatar do?

The avatar is more than an animated speech bubble. According to Heise, it can speak, see, and facially react to users – meaning it's not just outputting audio, but also perceiving and responding to your facial expressions and moods.

According to Golem, it speaks 97 languages. This should make it especially interesting for international companies that operate multilingual assistants or run corresponding customer service.

Who gets the avatar first?

The most important limitation: according to both sources, the feature is initially aimed at businesses. When and whether private users will be able to use the avatar does not emerge from the available reports.

Difference between the reports

Both sources report on the same announcement, but name the product slightly differently. Golem writes about "Gemini 3.8 Live", Heise only refers to "Gemini". Which designation is the official one does not emerge from the available reports.